End of training

README.md

@@ -23,7 +23,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.9456521739130435
+      value: 0.9565217391304348
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,8 +33,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/vit-msn-small](https://huggingface.co/facebook/vit-msn-small) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3181
-- Accuracy: 0.9457
+- Loss: 0.1189
+- Accuracy: 0.9565
 
 ## Model description
 

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 60.0,
     "eval_accuracy": 0.9565217391304348,
-    "eval_loss": 0.17970529198646545,
-    "eval_runtime": 0.3194,
-    "eval_samples_per_second": 288.014,
-    "eval_steps_per_second": 9.392,
+    "eval_loss": 0.11890643835067749,
+    "eval_runtime": 0.2938,
+    "eval_samples_per_second": 313.086,
+    "eval_steps_per_second": 10.209,
     "total_flos": 4.3676735454019584e+17,
-    "train_loss": 0.10428427308797836,
-    "train_runtime": 199.9604,
-    "train_samples_per_second": 111.622,
-    "train_steps_per_second": 0.9
+    "train_loss": 0.1091307305627399,
+    "train_runtime": 203.4457,
+    "train_samples_per_second": 109.71,
+    "train_steps_per_second": 0.885
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 60.0,
     "eval_accuracy": 0.9565217391304348,
-    "eval_loss": 0.17970529198646545,
-    "eval_runtime": 0.3194,
-    "eval_samples_per_second": 288.014,
-    "eval_steps_per_second": 9.392
+    "eval_loss": 0.11890643835067749,
+    "eval_runtime": 0.2938,
+    "eval_samples_per_second": 313.086,
+    "eval_steps_per_second": 10.209
 }

runs/Dec11_16-13-35_ae1aa77fe319/events.out.tfevents.1733933846.ae1aa77fe319.236.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:603a6f03df864d4e7b55674d1a10efcdeb4c35b59a869ab209a744df0e73d20d
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 60.0,
     "total_flos": 4.3676735454019584e+17,
-    "train_loss": 0.10428427308797836,
-    "train_runtime": 199.9604,
-    "train_samples_per_second": 111.622,
-    "train_steps_per_second": 0.9
+    "train_loss": 0.1091307305627399,
+    "train_runtime": 203.4457,
+    "train_samples_per_second": 109.71,
+    "train_steps_per_second": 0.885
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_metric": 0.9565217391304348,
-  "best_model_checkpoint": "vit-msn-small-wbc-blur-detector/checkpoint-60",
+  "best_model_checkpoint": "vit-msn-small-wbc-blur-detector/checkpoint-15",
   "epoch": 60.0,
   "eval_steps": 500,
   "global_step": 180,
@@ -10,678 +10,678 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5760869565217391,
-      "eval_loss": 0.6225717663764954,
-      "eval_runtime": 0.2813,
-      "eval_samples_per_second": 327.047,
-      "eval_steps_per_second": 10.665,
+      "eval_accuracy": 0.6304347826086957,
+      "eval_loss": 0.6141545176506042,
+      "eval_runtime": 0.3002,
+      "eval_samples_per_second": 306.507,
+      "eval_steps_per_second": 9.995,
       "step": 3
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8913043478260869,
-      "eval_loss": 0.41749870777130127,
-      "eval_runtime": 0.2698,
-      "eval_samples_per_second": 341.032,
-      "eval_steps_per_second": 11.121,
+      "eval_accuracy": 0.8695652173913043,
+      "eval_loss": 0.3853473365306854,
+      "eval_runtime": 0.2807,
+      "eval_samples_per_second": 327.721,
+      "eval_steps_per_second": 10.687,
       "step": 6
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8369565217391305,
-      "eval_loss": 0.359576016664505,
-      "eval_runtime": 0.2715,
-      "eval_samples_per_second": 338.881,
-      "eval_steps_per_second": 11.05,
+      "eval_accuracy": 0.8260869565217391,
+      "eval_loss": 0.40699997544288635,
+      "eval_runtime": 0.2954,
+      "eval_samples_per_second": 311.393,
+      "eval_steps_per_second": 10.154,
       "step": 9
     },
     {
       "epoch": 3.3333333333333335,
-      "grad_norm": 5.03557014465332,
+      "grad_norm": 6.74481725692749,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.5806,
+      "loss": 0.494,
       "step": 10
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9130434782608695,
-      "eval_loss": 0.2136119157075882,
-      "eval_runtime": 0.2933,
-      "eval_samples_per_second": 313.644,
-      "eval_steps_per_second": 10.228,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.1460711508989334,
+      "eval_runtime": 0.2916,
+      "eval_samples_per_second": 315.469,
+      "eval_steps_per_second": 10.287,
       "step": 12
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.18237723410129547,
-      "eval_runtime": 0.2735,
-      "eval_samples_per_second": 336.44,
-      "eval_steps_per_second": 10.971,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.11890643835067749,
+      "eval_runtime": 0.2956,
+      "eval_samples_per_second": 311.209,
+      "eval_steps_per_second": 10.148,
       "step": 15
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.9130434782608695,
-      "eval_loss": 0.21223501861095428,
-      "eval_runtime": 0.2858,
-      "eval_samples_per_second": 321.851,
-      "eval_steps_per_second": 10.495,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.15268591046333313,
+      "eval_runtime": 0.2919,
+      "eval_samples_per_second": 315.184,
+      "eval_steps_per_second": 10.278,
       "step": 18
     },
     {
       "epoch": 6.666666666666667,
-      "grad_norm": 29.32073402404785,
+      "grad_norm": 32.091392517089844,
       "learning_rate": 4.938271604938271e-05,
-      "loss": 0.2353,
+      "loss": 0.2024,
       "step": 20
     },
     {
       "epoch": 7.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2082703560590744,
-      "eval_runtime": 0.2935,
-      "eval_samples_per_second": 313.469,
-      "eval_steps_per_second": 10.222,
+      "eval_accuracy": 0.9021739130434783,
+      "eval_loss": 0.3323056697845459,
+      "eval_runtime": 0.2758,
+      "eval_samples_per_second": 333.524,
+      "eval_steps_per_second": 10.876,
       "step": 21
     },
     {
       "epoch": 8.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.1384642869234085,
-      "eval_runtime": 0.277,
-      "eval_samples_per_second": 332.147,
-      "eval_steps_per_second": 10.831,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.15201511979103088,
+      "eval_runtime": 0.2782,
+      "eval_samples_per_second": 330.705,
+      "eval_steps_per_second": 10.784,
       "step": 24
     },
     {
       "epoch": 9.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.15815076231956482,
-      "eval_runtime": 0.2722,
-      "eval_samples_per_second": 338.031,
-      "eval_steps_per_second": 11.023,
+      "eval_loss": 0.1572241336107254,
+      "eval_runtime": 0.2775,
+      "eval_samples_per_second": 331.58,
+      "eval_steps_per_second": 10.812,
       "step": 27
     },
     {
       "epoch": 10.0,
-      "grad_norm": 7.045929431915283,
+      "grad_norm": 10.290038108825684,
       "learning_rate": 4.62962962962963e-05,
-      "loss": 0.1214,
+      "loss": 0.1419,
       "step": 30
     },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.18020795285701752,
-      "eval_runtime": 0.3138,
-      "eval_samples_per_second": 293.183,
-      "eval_steps_per_second": 9.56,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.18135777115821838,
+      "eval_runtime": 0.2913,
+      "eval_samples_per_second": 315.861,
+      "eval_steps_per_second": 10.3,
       "step": 30
     },
     {
       "epoch": 11.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.1709580272436142,
-      "eval_runtime": 0.2875,
-      "eval_samples_per_second": 319.953,
-      "eval_steps_per_second": 10.433,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.17784903943538666,
+      "eval_runtime": 0.2749,
+      "eval_samples_per_second": 334.701,
+      "eval_steps_per_second": 10.914,
       "step": 33
     },
     {
       "epoch": 12.0,
       "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.2764737904071808,
-      "eval_runtime": 0.274,
-      "eval_samples_per_second": 335.737,
-      "eval_steps_per_second": 10.948,
+      "eval_loss": 0.15048673748970032,
+      "eval_runtime": 0.2937,
+      "eval_samples_per_second": 313.271,
+      "eval_steps_per_second": 10.215,
       "step": 36
     },
     {
       "epoch": 13.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.22897538542747498,
-      "eval_runtime": 0.2743,
-      "eval_samples_per_second": 335.388,
-      "eval_steps_per_second": 10.937,
+      "eval_loss": 0.1890527755022049,
+      "eval_runtime": 0.3016,
+      "eval_samples_per_second": 305.089,
+      "eval_steps_per_second": 9.949,
       "step": 39
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 4.235286235809326,
+      "grad_norm": 6.0901970863342285,
       "learning_rate": 4.3209876543209875e-05,
-      "loss": 0.0934,
+      "loss": 0.1053,
       "step": 40
     },
     {
       "epoch": 14.0,
-      "eval_accuracy": 0.9239130434782609,
-      "eval_loss": 0.31263482570648193,
-      "eval_runtime": 0.2763,
-      "eval_samples_per_second": 332.946,
-      "eval_steps_per_second": 10.857,
+      "eval_accuracy": 0.7934782608695652,
+      "eval_loss": 0.7274413108825684,
+      "eval_runtime": 0.272,
+      "eval_samples_per_second": 338.198,
+      "eval_steps_per_second": 11.028,
       "step": 42
     },
     {
       "epoch": 15.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.17815206944942474,
-      "eval_runtime": 0.3126,
-      "eval_samples_per_second": 294.294,
-      "eval_steps_per_second": 9.597,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.2668982148170471,
+      "eval_runtime": 0.2934,
+      "eval_samples_per_second": 313.551,
+      "eval_steps_per_second": 10.224,
       "step": 45
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 0.8804347826086957,
-      "eval_loss": 0.37992385029792786,
-      "eval_runtime": 0.2777,
-      "eval_samples_per_second": 331.253,
-      "eval_steps_per_second": 10.802,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.22397232055664062,
+      "eval_runtime": 0.3087,
+      "eval_samples_per_second": 298.059,
+      "eval_steps_per_second": 9.719,
       "step": 48
     },
     {
       "epoch": 16.666666666666668,
-      "grad_norm": 8.808466911315918,
+      "grad_norm": 7.414670467376709,
       "learning_rate": 4.012345679012346e-05,
-      "loss": 0.1525,
+      "loss": 0.3044,
       "step": 50
     },
     {
       "epoch": 17.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.19111250340938568,
-      "eval_runtime": 0.2894,
-      "eval_samples_per_second": 317.845,
-      "eval_steps_per_second": 10.365,
+      "eval_accuracy": 0.8913043478260869,
+      "eval_loss": 0.3497091829776764,
+      "eval_runtime": 0.2758,
+      "eval_samples_per_second": 333.612,
+      "eval_steps_per_second": 10.879,
       "step": 51
     },
     {
       "epoch": 18.0,
       "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.22915299236774445,
-      "eval_runtime": 0.2754,
-      "eval_samples_per_second": 334.072,
-      "eval_steps_per_second": 10.894,
+      "eval_loss": 0.2208346724510193,
+      "eval_runtime": 0.274,
+      "eval_samples_per_second": 335.712,
+      "eval_steps_per_second": 10.947,
       "step": 54
     },
     {
       "epoch": 19.0,
-      "eval_accuracy": 0.9130434782608695,
-      "eval_loss": 0.3240966498851776,
-      "eval_runtime": 0.29,
-      "eval_samples_per_second": 317.233,
-      "eval_steps_per_second": 10.345,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.17329861223697662,
+      "eval_runtime": 0.285,
+      "eval_samples_per_second": 322.844,
+      "eval_steps_per_second": 10.528,
       "step": 57
     },
     {
       "epoch": 20.0,
-      "grad_norm": 12.596190452575684,
+      "grad_norm": 5.56092643737793,
       "learning_rate": 3.7037037037037037e-05,
-      "loss": 0.1551,
+      "loss": 0.151,
       "step": 60
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.9565217391304348,
-      "eval_loss": 0.17970529198646545,
-      "eval_runtime": 0.2984,
-      "eval_samples_per_second": 308.312,
-      "eval_steps_per_second": 10.054,
+      "eval_accuracy": 0.9239130434782609,
+      "eval_loss": 0.20381057262420654,
+      "eval_runtime": 0.2821,
+      "eval_samples_per_second": 326.07,
+      "eval_steps_per_second": 10.633,
       "step": 60
     },
     {
       "epoch": 21.0,
-      "eval_accuracy": 0.9239130434782609,
-      "eval_loss": 0.2115555703639984,
-      "eval_runtime": 0.2749,
-      "eval_samples_per_second": 334.611,
-      "eval_steps_per_second": 10.911,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.12818782031536102,
+      "eval_runtime": 0.2765,
+      "eval_samples_per_second": 332.708,
+      "eval_steps_per_second": 10.849,
       "step": 63
     },
     {
       "epoch": 22.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.2774827480316162,
-      "eval_runtime": 0.2726,
-      "eval_samples_per_second": 337.516,
-      "eval_steps_per_second": 11.006,
+      "eval_accuracy": 0.9239130434782609,
+      "eval_loss": 0.3231411874294281,
+      "eval_runtime": 0.3046,
+      "eval_samples_per_second": 302.028,
+      "eval_steps_per_second": 9.849,
       "step": 66
     },
     {
       "epoch": 23.0,
-      "eval_accuracy": 0.9239130434782609,
-      "eval_loss": 0.28300002217292786,
-      "eval_runtime": 0.2723,
-      "eval_samples_per_second": 337.81,
-      "eval_steps_per_second": 11.016,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.15651515126228333,
+      "eval_runtime": 0.296,
+      "eval_samples_per_second": 310.773,
+      "eval_steps_per_second": 10.134,
       "step": 69
     },
     {
       "epoch": 23.333333333333332,
-      "grad_norm": 3.0055344104766846,
+      "grad_norm": 6.024500370025635,
       "learning_rate": 3.395061728395062e-05,
-      "loss": 0.0701,
+      "loss": 0.0875,
       "step": 70
     },
     {
       "epoch": 24.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.33308491110801697,
-      "eval_runtime": 0.2939,
-      "eval_samples_per_second": 313.029,
-      "eval_steps_per_second": 10.207,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.19810304045677185,
+      "eval_runtime": 0.3099,
+      "eval_samples_per_second": 296.864,
+      "eval_steps_per_second": 9.68,
       "step": 72
     },
     {
       "epoch": 25.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.27504873275756836,
-      "eval_runtime": 0.2721,
-      "eval_samples_per_second": 338.147,
-      "eval_steps_per_second": 11.027,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.19737455248832703,
+      "eval_runtime": 0.2724,
+      "eval_samples_per_second": 337.687,
+      "eval_steps_per_second": 11.012,
       "step": 75
     },
     {
       "epoch": 26.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2002851665019989,
-      "eval_runtime": 0.3046,
-      "eval_samples_per_second": 302.042,
-      "eval_steps_per_second": 9.849,
+      "eval_loss": 0.20447766780853271,
+      "eval_runtime": 0.2918,
+      "eval_samples_per_second": 315.301,
+      "eval_steps_per_second": 10.282,
       "step": 78
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 3.5597548484802246,
+      "grad_norm": 4.34276008605957,
       "learning_rate": 3.08641975308642e-05,
-      "loss": 0.0733,
+      "loss": 0.0851,
       "step": 80
     },
     {
       "epoch": 27.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2492789775133133,
-      "eval_runtime": 0.2789,
-      "eval_samples_per_second": 329.913,
-      "eval_steps_per_second": 10.758,
+      "eval_loss": 0.1840977817773819,
+      "eval_runtime": 0.2713,
+      "eval_samples_per_second": 339.118,
+      "eval_steps_per_second": 11.058,
       "step": 81
     },
     {
       "epoch": 28.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2807539105415344,
-      "eval_runtime": 0.2756,
-      "eval_samples_per_second": 333.87,
-      "eval_steps_per_second": 10.887,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.20613481104373932,
+      "eval_runtime": 0.2718,
+      "eval_samples_per_second": 338.491,
+      "eval_steps_per_second": 11.038,
       "step": 84
     },
     {
       "epoch": 29.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2663654088973999,
-      "eval_runtime": 0.2778,
-      "eval_samples_per_second": 331.156,
-      "eval_steps_per_second": 10.799,
+      "eval_loss": 0.20765484869480133,
+      "eval_runtime": 0.2926,
+      "eval_samples_per_second": 314.467,
+      "eval_steps_per_second": 10.254,
       "step": 87
     },
     {
       "epoch": 30.0,
-      "grad_norm": 2.488931179046631,
+      "grad_norm": 2.4659526348114014,
       "learning_rate": 2.777777777777778e-05,
-      "loss": 0.0494,
+      "loss": 0.046,
       "step": 90
     },
     {
       "epoch": 30.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.36214956641197205,
-      "eval_runtime": 0.2768,
-      "eval_samples_per_second": 332.394,
-      "eval_steps_per_second": 10.839,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.21993966400623322,
+      "eval_runtime": 0.2942,
+      "eval_samples_per_second": 312.716,
+      "eval_steps_per_second": 10.197,
       "step": 90
     },
     {
       "epoch": 31.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2520482838153839,
-      "eval_runtime": 0.2749,
-      "eval_samples_per_second": 334.646,
-      "eval_steps_per_second": 10.912,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.2038496881723404,
+      "eval_runtime": 0.3027,
+      "eval_samples_per_second": 303.903,
+      "eval_steps_per_second": 9.91,
       "step": 93
     },
     {
       "epoch": 32.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2721933424472809,
-      "eval_runtime": 0.2755,
-      "eval_samples_per_second": 333.902,
-      "eval_steps_per_second": 10.888,
+      "eval_loss": 0.20770704746246338,
+      "eval_runtime": 0.281,
+      "eval_samples_per_second": 327.359,
+      "eval_steps_per_second": 10.675,
       "step": 96
     },
     {
       "epoch": 33.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.27847108244895935,
-      "eval_runtime": 0.277,
-      "eval_samples_per_second": 332.097,
-      "eval_steps_per_second": 10.829,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.18766026198863983,
+      "eval_runtime": 0.2757,
+      "eval_samples_per_second": 333.675,
+      "eval_steps_per_second": 10.881,
       "step": 99
     },
     {
       "epoch": 33.333333333333336,
-      "grad_norm": 13.367582321166992,
+      "grad_norm": 1.9909802675247192,
       "learning_rate": 2.4691358024691357e-05,
-      "loss": 0.0806,
+      "loss": 0.0533,
       "step": 100
     },
     {
       "epoch": 34.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2530038356781006,
-      "eval_runtime": 0.2765,
-      "eval_samples_per_second": 332.674,
-      "eval_steps_per_second": 10.848,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.2383040189743042,
+      "eval_runtime": 0.2759,
+      "eval_samples_per_second": 333.465,
+      "eval_steps_per_second": 10.874,
       "step": 102
     },
     {
       "epoch": 35.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.22303055226802826,
-      "eval_runtime": 0.2793,
-      "eval_samples_per_second": 329.409,
-      "eval_steps_per_second": 10.742,
+      "eval_accuracy": 0.9239130434782609,
+      "eval_loss": 0.25708499550819397,
+      "eval_runtime": 0.2913,
+      "eval_samples_per_second": 315.812,
+      "eval_steps_per_second": 10.298,
       "step": 105
     },
     {
       "epoch": 36.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.2482990175485611,
-      "eval_runtime": 0.2776,
-      "eval_samples_per_second": 331.464,
-      "eval_steps_per_second": 10.809,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.23297645151615143,
+      "eval_runtime": 0.2763,
+      "eval_samples_per_second": 332.955,
+      "eval_steps_per_second": 10.857,
       "step": 108
     },
     {
       "epoch": 36.666666666666664,
-      "grad_norm": 7.69065523147583,
+      "grad_norm": 9.732784271240234,
       "learning_rate": 2.1604938271604937e-05,
-      "loss": 0.0324,
+      "loss": 0.0451,
       "step": 110
     },
     {
       "epoch": 37.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.29446083307266235,
-      "eval_runtime": 0.2954,
-      "eval_samples_per_second": 311.393,
-      "eval_steps_per_second": 10.154,
+      "eval_loss": 0.24198591709136963,
+      "eval_runtime": 0.2748,
+      "eval_samples_per_second": 334.8,
+      "eval_steps_per_second": 10.917,
       "step": 111
     },
     {
       "epoch": 38.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.3244095742702484,
-      "eval_runtime": 0.2815,
-      "eval_samples_per_second": 326.826,
-      "eval_steps_per_second": 10.657,
+      "eval_accuracy": 0.9239130434782609,
+      "eval_loss": 0.2881980240345001,
+      "eval_runtime": 0.291,
+      "eval_samples_per_second": 316.1,
+      "eval_steps_per_second": 10.308,
       "step": 114
     },
     {
       "epoch": 39.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.3302266597747803,
-      "eval_runtime": 0.2668,
-      "eval_samples_per_second": 344.764,
-      "eval_steps_per_second": 11.242,
+      "eval_loss": 0.23858585953712463,
+      "eval_runtime": 0.2975,
+      "eval_samples_per_second": 309.204,
+      "eval_steps_per_second": 10.083,
       "step": 117
     },
     {
       "epoch": 40.0,
-      "grad_norm": 8.54801082611084,
+      "grad_norm": 1.4663212299346924,
       "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.0435,
+      "loss": 0.0401,
       "step": 120
     },
     {
       "epoch": 40.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.331495463848114,
-      "eval_runtime": 0.275,
-      "eval_samples_per_second": 334.557,
-      "eval_steps_per_second": 10.909,
+      "eval_accuracy": 0.9347826086956522,
+      "eval_loss": 0.2512564957141876,
+      "eval_runtime": 0.2721,
+      "eval_samples_per_second": 338.098,
+      "eval_steps_per_second": 11.025,
       "step": 120
     },
     {
       "epoch": 41.0,
       "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.303882896900177,
-      "eval_runtime": 0.2881,
-      "eval_samples_per_second": 319.318,
-      "eval_steps_per_second": 10.413,
+      "eval_loss": 0.2671690285205841,
+      "eval_runtime": 0.2823,
+      "eval_samples_per_second": 325.908,
+      "eval_steps_per_second": 10.627,
       "step": 123
     },
     {
       "epoch": 42.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.36054134368896484,
-      "eval_runtime": 0.2765,
-      "eval_samples_per_second": 332.764,
-      "eval_steps_per_second": 10.851,
+      "eval_loss": 0.2950490117073059,
+      "eval_runtime": 0.2748,
+      "eval_samples_per_second": 334.771,
+      "eval_steps_per_second": 10.916,
       "step": 126
     },
     {
       "epoch": 43.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3643103539943695,
-      "eval_runtime": 0.2772,
-      "eval_samples_per_second": 331.854,
-      "eval_steps_per_second": 10.821,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.32324346899986267,
+      "eval_runtime": 0.2744,
+      "eval_samples_per_second": 335.279,
+      "eval_steps_per_second": 10.933,
       "step": 129
     },
     {
       "epoch": 43.333333333333336,
-      "grad_norm": 5.6320648193359375,
+      "grad_norm": 6.655643463134766,
       "learning_rate": 1.54320987654321e-05,
-      "loss": 0.0325,
+      "loss": 0.0329,
       "step": 130
     },
     {
       "epoch": 44.0,
-      "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.34682697057724,
-      "eval_runtime": 0.2993,
-      "eval_samples_per_second": 307.404,
-      "eval_steps_per_second": 10.024,
+      "eval_accuracy": 0.9239130434782609,
+      "eval_loss": 0.3711928427219391,
+      "eval_runtime": 0.2829,
+      "eval_samples_per_second": 325.258,
+      "eval_steps_per_second": 10.606,
       "step": 132
     },
     {
       "epoch": 45.0,
       "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.33276745676994324,
-      "eval_runtime": 0.2866,
-      "eval_samples_per_second": 321.025,
-      "eval_steps_per_second": 10.468,
+      "eval_loss": 0.35285505652427673,
+      "eval_runtime": 0.2761,
+      "eval_samples_per_second": 333.224,
+      "eval_steps_per_second": 10.866,
       "step": 135
     },
     {
       "epoch": 46.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3260520398616791,
-      "eval_runtime": 0.2798,
-      "eval_samples_per_second": 328.753,
-      "eval_steps_per_second": 10.72,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.29050472378730774,
+      "eval_runtime": 0.292,
+      "eval_samples_per_second": 315.053,
+      "eval_steps_per_second": 10.273,
       "step": 138
     },
     {
       "epoch": 46.666666666666664,
-      "grad_norm": 2.9780068397521973,
+      "grad_norm": 5.307190895080566,
       "learning_rate": 1.2345679012345678e-05,
-      "loss": 0.0541,
+      "loss": 0.0519,
       "step": 140
     },
     {
       "epoch": 47.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.34095829725265503,
-      "eval_runtime": 0.2824,
-      "eval_samples_per_second": 325.795,
-      "eval_steps_per_second": 10.624,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.26701638102531433,
+      "eval_runtime": 0.3064,
+      "eval_samples_per_second": 300.308,
+      "eval_steps_per_second": 9.793,
       "step": 141
     },
     {
       "epoch": 48.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3532644212245941,
-      "eval_runtime": 0.2895,
-      "eval_samples_per_second": 317.751,
-      "eval_steps_per_second": 10.361,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.2628593444824219,
+      "eval_runtime": 0.2788,
+      "eval_samples_per_second": 330.007,
+      "eval_steps_per_second": 10.761,
       "step": 144
     },
     {
       "epoch": 49.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3564309775829315,
-      "eval_runtime": 0.3048,
-      "eval_samples_per_second": 301.792,
-      "eval_steps_per_second": 9.841,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.2760757505893707,
+      "eval_runtime": 0.2738,
+      "eval_samples_per_second": 336.03,
+      "eval_steps_per_second": 10.957,
       "step": 147
     },
     {
       "epoch": 50.0,
-      "grad_norm": 3.5734236240386963,
+      "grad_norm": 4.582086563110352,
       "learning_rate": 9.259259259259259e-06,
-      "loss": 0.022,
+      "loss": 0.0281,
       "step": 150
     },
     {
       "epoch": 50.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3557578921318054,
-      "eval_runtime": 0.2742,
-      "eval_samples_per_second": 335.507,
-      "eval_steps_per_second": 10.94,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.3040381669998169,
+      "eval_runtime": 0.2766,
+      "eval_samples_per_second": 332.655,
+      "eval_steps_per_second": 10.847,
       "step": 150
     },
     {
       "epoch": 51.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.34129852056503296,
-      "eval_runtime": 0.311,
-      "eval_samples_per_second": 295.815,
-      "eval_steps_per_second": 9.646,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.31911370158195496,
+      "eval_runtime": 0.2968,
+      "eval_samples_per_second": 309.998,
+      "eval_steps_per_second": 10.109,
       "step": 153
     },
     {
       "epoch": 52.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.33716654777526855,
-      "eval_runtime": 0.2804,
-      "eval_samples_per_second": 328.129,
-      "eval_steps_per_second": 10.7,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.32143697142601013,
+      "eval_runtime": 0.3072,
+      "eval_samples_per_second": 299.499,
+      "eval_steps_per_second": 9.766,
       "step": 156
     },
     {
       "epoch": 53.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.33803921937942505,
-      "eval_runtime": 0.301,
-      "eval_samples_per_second": 305.697,
-      "eval_steps_per_second": 9.968,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.31315502524375916,
+      "eval_runtime": 0.3028,
+      "eval_samples_per_second": 303.834,
+      "eval_steps_per_second": 9.908,
       "step": 159
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 8.637144088745117,
+      "grad_norm": 4.059518814086914,
       "learning_rate": 6.172839506172839e-06,
-      "loss": 0.0226,
+      "loss": 0.028,
       "step": 160
     },
     {
       "epoch": 54.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.32888463139533997,
-      "eval_runtime": 0.2786,
-      "eval_samples_per_second": 330.184,
-      "eval_steps_per_second": 10.767,
+      "eval_loss": 0.31154391169548035,
+      "eval_runtime": 0.2938,
+      "eval_samples_per_second": 313.128,
+      "eval_steps_per_second": 10.211,
       "step": 162
     },
     {
       "epoch": 55.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3256872296333313,
-      "eval_runtime": 0.2749,
-      "eval_samples_per_second": 334.655,
-      "eval_steps_per_second": 10.913,
+      "eval_accuracy": 0.9565217391304348,
+      "eval_loss": 0.31155669689178467,
+      "eval_runtime": 0.2755,
+      "eval_samples_per_second": 333.966,
+      "eval_steps_per_second": 10.89,
       "step": 165
     },
     {
       "epoch": 56.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.32627877593040466,
-      "eval_runtime": 0.3023,
-      "eval_samples_per_second": 304.338,
-      "eval_steps_per_second": 9.924,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.3224737346172333,
+      "eval_runtime": 0.302,
+      "eval_samples_per_second": 304.679,
+      "eval_steps_per_second": 9.935,
       "step": 168
     },
     {
       "epoch": 56.666666666666664,
-      "grad_norm": 4.384215831756592,
+      "grad_norm": 9.457426071166992,
       "learning_rate": 3.0864197530864196e-06,
-      "loss": 0.0358,
+      "loss": 0.0361,
       "step": 170
     },
     {
       "epoch": 57.0,
-      "eval_accuracy": 0.9347826086956522,
-      "eval_loss": 0.3316076099872589,
-      "eval_runtime": 0.3111,
-      "eval_samples_per_second": 295.734,
-      "eval_steps_per_second": 9.644,
+      "eval_accuracy": 0.9456521739130435,
+      "eval_loss": 0.3235200047492981,
+      "eval_runtime": 0.2841,
+      "eval_samples_per_second": 323.798,
+      "eval_steps_per_second": 10.559,
       "step": 171
     },
     {
       "epoch": 58.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.3367496728897095,
-      "eval_runtime": 0.2975,
-      "eval_samples_per_second": 309.229,
-      "eval_steps_per_second": 10.084,
+      "eval_loss": 0.3200394809246063,
+      "eval_runtime": 0.2914,
+      "eval_samples_per_second": 315.666,
+      "eval_steps_per_second": 10.293,
       "step": 174
     },
     {
       "epoch": 59.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.33811014890670776,
-      "eval_runtime": 0.2965,
-      "eval_samples_per_second": 310.296,
-      "eval_steps_per_second": 10.118,
+      "eval_loss": 0.318322092294693,
+      "eval_runtime": 0.2746,
+      "eval_samples_per_second": 335.002,
+      "eval_steps_per_second": 10.924,
       "step": 177
     },
     {
       "epoch": 60.0,
-      "grad_norm": 6.692429065704346,
+      "grad_norm": 9.799005508422852,
       "learning_rate": 0.0,
-      "loss": 0.0225,
+      "loss": 0.0312,
       "step": 180
     },
     {
       "epoch": 60.0,
       "eval_accuracy": 0.9456521739130435,
-      "eval_loss": 0.33829161524772644,
-      "eval_runtime": 0.2793,
-      "eval_samples_per_second": 329.443,
-      "eval_steps_per_second": 10.743,
+      "eval_loss": 0.31814736127853394,
+      "eval_runtime": 0.2764,
+      "eval_samples_per_second": 332.886,
+      "eval_steps_per_second": 10.855,
       "step": 180
     },
     {
       "epoch": 60.0,
       "step": 180,
       "total_flos": 4.3676735454019584e+17,
-      "train_loss": 0.10428427308797836,
-      "train_runtime": 199.9604,
-      "train_samples_per_second": 111.622,
-      "train_steps_per_second": 0.9
+      "train_loss": 0.1091307305627399,
+      "train_runtime": 203.4457,
+      "train_samples_per_second": 109.71,
+      "train_steps_per_second": 0.885
     }
   ],
   "logging_steps": 10,